Improving Heterogeneous Workload Performance in Server Virtualization Based on User Behaviors
Dulyawit Prangchumpol, Siripun Sanguansintukul, Panjai Tantasanawong · Journal of Convergence Information Technology · 2012
Many organizations emphasize the reduction of power consumption in their IT enterprises. Virtualization techniques are gaining attention in recent years. Server virtualization can help to consolidate servers, improve hardware utilization and reduce the consumption of power and physical space in the data center. However, management of heterogeneous workloads in this system becomes a challenge. This research proposes a new concept for managing workloads based on user behavior. The experiment is categorized into two parts. First, data mining technique is used to explore the trend of user behaviors in each server service. The preliminary results showed that user behaviors are different in each type of service workload and time. Next, the prediction focuses on hardware resources, CPU and memory based on real behaviors are presented. Different prediction algorithms such as a simple exponential smoothing, a double exponential smoothing (Holt’s method) and a winters’ threeparameter trend and seasonal method are then applied.